IRC3 Regulates Mitochondrial Translation in Response to Metabolic Cues in Saccharomyces cerevisiae

ABSTRACT

Mitochondrial oxidative phosphorylation (OXPHOS) enzymes have a dual genetic origin. Mechanisms regulating the expression of nucleus-encoded OXPHOS subunits in response to metabolic cues (glucose versus glycerol) are well understood, while the regulation of mitochondrially encoded OXPHOS subunits is poorly defined. Here, we show that IRC3, a DEAD/H box helicase gene, previously implicated in mitochondrial DNA maintenance, is central to integrating metabolic cues with mitochondrial translation. Irc3 associates with mitochondrial small ribosomal subunits in cells consistent with its role in regulating translation elongation based on the Arg8^m reporter system. IRC3-deleted cells retained mitochondrial DNA despite a growth defect on glycerol plates. Glucose-grown Δirc3ρ⁺ and irc3 temperature-sensitive cells at 37°C have reduced translation rates from the majority of mRNAs. In contrast, when galactose was the carbon source, a reduction in mitochondrial translation was observed predominantly from Cox1 mRNA in Δirc3ρ⁺ cells but no defect was observed in irc3 temperature-sensitive cells, at 37°C. In support of a model whereby IRC3 responds to metabolic cues to regulate mitochondrial translation, Δirc3 suppressor strains isolated for restoration of growth on glycerol medium restore mitochondrial protein synthesis differentially in the presence of glucose versus glycerol.

INTRODUCTION

Mitochondria, which are best known as the powerhouse of the cell, require coordinated gene expression of two spatially distinct genetic materials. Mitochondria are essential for an organism’s viability and normal physiology, and any disruption in its functioning leads to a myriad of cellular defects, including cancer (1–3). Saccharomyces cerevisiae has been an invaluable system for deciphering mitochondrial function, due to its ability to survive without respiration as well as mitochondrial DNA (mtDNA), permitting the characterization of mutants that impair mitochondrial functioning. In Saccharomyces cerevisiae, mtDNA encodes eight proteins, seven of which are transmembrane proteins that are essential components of oxidative phosphorylation (OXPHOS) machinery and one, soluble protein Var1, that is an essential component of mitochondrial small ribosomal subunits. In addition to these eight protein coding genes, mtDNA also encodes rRNAs (15S and 21S) and a complete set of tRNAs required for gene expression. The remainder of the proteins that make up the OXPHOS subunits, factors required for mitochondrial transcription and translation, including mitochondrial ribosomal proteins, are encoded by the nuclear genome, which is translated in the cytosol and imported into the mitochondria (4).

Translation of mitochondrial mRNA, in addition to general translation factors such as Tuf1 (EF-Tu) and Mef1 and Mef2 (EF-G), require membrane-bound mRNA-specific translation activators (5–11). In the absence of either Shine-Dalgarno sequences or a 5′ cap on mitochondrial transcripts, these mRNA-specific translation activators recognize the 5′ untranslated region (UTR) of mRNA to localize them to the mitochondrial inner membrane, where they aid in the loading of membrane-bound ribosomes to initiate mitochondrial translation (12–18). In fact, each mitochondrial mRNA has its specific set of translation activators (19, 20). Interestingly, altered levels or activities of these translation activators are thought to allow yeast cells to sense metabolic cues. For yeast cells, transfer from fermentable to nonfermentable growth leads to enhanced OXPHOS expression and activity. This is brought about by enhanced expression of nucleus-encoded OXPHOS subunits and mitochondrial mRNA-specific translation factors, which upregulates translation of mitochondrially encoded OXPHOS subunits (21–23). Another way mitochondrial translation responds to alteration in metabolic cues is by using different mitochondrial translation factors under different conditions. For example, the translation from COX1 mRNA is controlled by PET309, MSS51, and MAM33 (24–29). Mam33 is required for the translation of Cox1 at the basal level in glucose-grown yeast cells along with Pet309p and Mss51p. However, its function is dispensable in cells adapted for growth in glycerol, where COX1 mRNA translation is regulated primarily by Mss51p and Pet309p (29).

In addition to translation factors, optimal mitochondrial gene expression requires the essential function of RNA helicase, among other accessory proteins. They regulate every aspect of mitochondrial RNA metabolism, including RNA splicing, mRNA turnover/surveillance, translation, and ribosome biogenesis (30–32). RNA helicases utilize energy released from ATP hydrolysis to promote rearrangements required by nascent RNA molecules to arrive at a mature structure (33–35). These mature structures are achieved by the action of helicases to unwind double-stranded RNA (dsRNA), anneal single-stranded RNA (ssRNA), modify RNA-DNA hybrids, and displace proteins bound to RNA (36–38). In Saccharomyces cerevisiae, the nuclear genome encodes four mitochondrial helicases which belong to the SFII family of NTP-dependent remodelers containing a DExH/D motif that regulates different aspects of mitochondrial function. MSS116 has been shown to play a role in transcription, mRNA splicing, and translation (39–42), while MRH4 and SUV3 are known to be involved in ribosome biogenesis and RNA editing/turnover, respectively (43–46). The fourth mitochondrial helicase gene, IRC3, has been implicated in mtDNA recombination and repair (47, 48). However, these studies were carried out in strains deleted for IRC3, in which the loss of mtDNA could be either due to a direct role of Irc3 in regulating mtDNA recombination/repair or to an indirect consequence of Irc3’s role in an essential process involving RNA molecules, such as transcription and translation.

In the present study, we have shown that IRC3 is a central regulator of mitochondrial translation. Interestingly, Irc3p regulates mitochondrial translation distinctly in cells grown in a carbon source that maintains OXPHOS at basal (glucose) versus elevated (galactose/glycerol) levels. Glucose-grown Δirc3ρ⁺ cells were defective for overall mitochondrial protein synthesis in addition to reducing the accumulation of mitochondrial transcripts. In contrast, glucose-grown cells harboring irc3 temperature-sensitive mutants have reduced rates of translation, without having any consequence on the levels of mRNA transcripts or assembled mitochondrial ribosomal subunits, when shifted to the nonpermissive temperature. Interestingly, for galactose-grown Δirc3ρ⁺ cells and cells harboring irc3 temperature-sensitive mutants, a reduction in mitochondrial translation from Cox1 mRNA was observed in Δirc3ρ⁺ cells but not in cells harboring irc3 temperature-sensitive mutants at the nonpermissive temperature. Consistent with a role in translation regulation, Irc3 cofractionates with small ribosomal subunits when cells are grown in either glucose or glycerol. Importantly, we have shown that IRC3 is required for translation elongation in glucose-grown cells. Consistent with a hypothesis whereby the mechanisms of action and targets of IRC3 are different under conditions that maintain mitochondrial function at a basal level and those that have higher mitochondrial function, we were able to isolate Δirc3ρ⁺suppressor strains that restored mitochondrial translation to different degrees when the cells were grown in glucose versus glycerol.

RESULTS

IRC3 is essential for normal respiratory growth.

To determine whether IRC3 regulates an essential mitochondrial function in cells, we sporulated a heterozygous diploid Δirc3 strain generated in the lab. Although the numbers of colonies formed on glucose plates in freshly germinated Δirc3 or IRC3 spores were similar, Δirc3 spores formed two types of colonies which could be differentiated based on colony size. On glycerol plates, a lower number of Δirc3 spores formed CFUs than did IRC3 spores, although the colony sizes were comparable. In fact, the numbers of large colonies formed on glucose plates were equivalent to the numbers of colonies formed on glycerol plates (Fig. 1A). To minimize the effect of an unlinked mutation on the inability of Δirc3 spores to utilize glycerol, Δirc3/IRC3 cells episomally expressing wild-type IRC3 linked to URA3 were sporulated. The Δirc3 spores episomally expressing the IRC3 allele linked to URA3 were backcrossed six times as described in Material and Methods. The reduced ability of freshly generated Δirc3 cells (colonies obtained on 5FOAD [0.67% yeast nitrogen base without amino acids containing 0.1% 5-fluoro-orotic acid and 2% glucose]) to utilize glycerol as the sole carbon source was a fraction of the total viable cells in comparison to that of Δirc3 cells ectopically expressing IRC3 when cultured either in glucose, galactose, or glycerol (see Fig. S1A in the supplemental material). This indicated that ectopic expression of IRC3 was able to complement the glycerol growth defect. All further experiments were carried out using this six-time-backcrossed strain (KDY 1146). Furthermore, the percentage of total viable cells that were able to utilize glycerol as the sole carbon source was estimated in Δirc3 cells expressing either wild-type IRC3 or vector upon subculturing in glucose for 0 h and 24 h. We observed that 10% of progeny cells were competent for cellular respiration in a population of Δirc3 cells in comparison to 90% of wild-type cells upon subculturing in glucose or glycerol for 24 h (Fig. 1B; Fig. S1B). Thus, taken together, this indicates that IRC3 is essential for optimal mitochondrial function. Mitochondrial DNA integrity has been reported to be compromised in yeast cells defective for essential processes related to mitochondrial DNA maintenance and gene expression. Measurement of timing of the loss of ability to utilize growth on glycerol versus loss of mtDNA in mutant cells is often indicative of the primary role played by the wild-type copy of the mutant gene product (43, 49, 50). Thus, we determined the status of mtDNA (rho or ρ) in a population of haploid Δirc3 cells containing either empty vector or episomally expressing IRC3 upon subculturing in glucose medium for 0 h and 24 h at 30°C. Aliquots of cells at each time point were crossed with wild-type ρ⁰ cells of the opposite mating type. The ability of the diploids generated to utilize glycerol is indicative of the presence of mitochondrial DNA in the parental haploid Δirc3 cells. Approximately 90 to 100% of cells episomally expressing a wild-type copy of IRC3 retained mitochondrial DNA upon subculturing in glucose. In Δirc3 cells, approximately 85% of the cells retained mtDNA upon subculturing in glucose for 24 h (Fig. 1C). However, further incubation of Δirc3 cells led to loss of mitochondrial DNA, such that only 20% of cells retained mitochondrial DNA upon subculturing in glucose for 72 h (Fig. S1C). Further, to quantify mtDNA and analyze its stability in Δirc3 cells expressing either wild-type IRC3 or vector, upon subculturing them in glucose for 0 h and 24 h, we measured the mtDNA copy number with respect to the nuclear DNA (nDNA) copy number as described in Material and Methods. Ratios of mtDNA to nDNA were similar in cells deleted for IRC3 and in wild-type cells even after subculturing in glucose for 24 h (Fig. S1D). Thus, all subsequent experiments using Δirc3ρ⁺ cells were carried out under culture conditions such that cells retained mitochondrial DNA. Taken together, our results indicate that loss of growth on glycerol precedes loss of mitochondrial DNA in Δirc3 cells. This is suggestive of a defect in an essential process within the mitochondria in Δirc3 cells such as those linked to mitochondrial gene expression as the primary cause of loss of cellular respiration and loss of mitochondrial DNA as a secondary consequence. Similar phenotypes have been reported for numerous mutant genes that are defective in mitochondrial gene expression (43, 49, 50).

FIG 1.

Loss of growth on glycerol precedes loss of mtDNA in Δirc3ρ⁺ cells due to aberrant gene expression. (A) Tenfold serial dilutions of freshly germinated spores of wild-type and Δirc3 cells spotted on glucose or glycerol plates are shown. (B) Percentages of viable Δirc3 progeny cells expressing either an episomal copy of IRC3 or empty vector that can utilize glycerol as the sole carbon source at the indicated time points upon subculturing in glucose. Three independent colonies of wild-type and Δirc3 cells were used to calculate the % YPG-positive cells. Five independent replicates of this experiment were performed. (C) Percentages of mtDNA present in Δirc3 progeny expressing either an episomal copy of IRC3 or vector at the indicated time after subculture in glucose. Three independent colonies of wild-type and Δirc3 cells were used to calculate the % ρ⁺ cells. Five independent replicates of this experiment were performed. (D) Transcript levels of mitochondrial genes were assayed in Δirc3 cells expressing either an episomal copy of IRC3 or vector grown in glucose, galactose, and glycerol. 21S rRNA and 15S rRNA levels were detected by EtBr staining. Samples were normalized for total RNA. Multiple independent biological and technical replicates were carried out, and representative images are shown. (E) Upper panel, schematic representation of exons and introns in the COX1 and COB transcripts. Primers subsequently used to determine the splice variant in COX1 and COB transcripts are indicated by arrows. Lower panel, levels of transcripts of pre-mRNA of COX1 (COX1 aI5β, product of primers I6 F.P and 6B3) and COB ligated exon (Cytb LE, b14F.P, and b14R.P). (F) Mitochondrial protein synthesis was measured in Δirc3 cells expressing either an episomal copy of IRC3 or vector cultured in either glucose or galactose at 30°C. Labeling reaction was carried out by incorporation of ³⁵S-labeled methionine and cysteine in the presence of cycloheximide to inhibit cytosolic protein synthesis at 30°C for 30 min. Mitochondria from labeled cells were isolated, and equivalent amounts of mitochondrial proteins were separated on a 17.5% SDS-PAGE gel and transferred to a nitrocellulose membrane. Radiolabeled proteins were visualized by phosphorimaging. Shown at the bottom is a Coomassie blue-stained gel of radiolabeled mitochondrial protein extracts. Percentages of mitochondrial DNA present in the respective strains are indicated at the bottom of the gels in panels D to F.

IRC3 is essential for mitochondrial gene expression.

IRC3 is a putative helicase of the SFII family, a majority of which have been known to act on RNA molecules (51, 52). Thus, IRC3 could potentially be involved in transcription, splicing, RNA maturation and turnover, ribosome assembly, or translation to regulate mitochondrial gene expression. Mitochondrial transcript levels were examined in Δirc3ρ⁺ cells grown in either glucose, galactose, or glycerol by performing reverse transcriptase PCR as described in Materials and Methods. Transcript levels of COX1, ATP6, and ATP8 were reduced in Δirc3ρ⁺ cells in comparison to wild-type cells, although steady-state levels of COB, COX2, COX3, and ATP9 mRNA remained relatively unchanged (Fig. 1D). These observations were consistent irrespective of the carbon sources used for cell culture (Fig. 1D). To determine whether the reduced transcript levels in cells deleted for IRC3 is a consequence of improper intron splicing, we analyzed Δirc3ρ⁺ cells for efficient splicing of COX1 and COB transcripts. To check the accumulation of transcripts where the a15β intron is retained in cDNAs synthesized from COX1 transcripts, primers used for reverse transcriptase PCR were designed between a15β and exon7 (I6F.P and 6B3). To amplify the ligated exon from synthesized cDNA for COB transcripts, primers were designed from exonic regions upstream and downstream of b14 (b14F.P and b14R.P) introns (Fig. 1E, top). In comparison to wild-type cells, the levels of unspliced mRNA, i.e., COX1 a15β, did not accumulate in Δirc3ρ⁺ cells (Fig. 1E). The levels of COB ligated exon were similar to transcript levels of COB (Fig. 1E and D). In addition, we found that Δirc3 cells harboring intronless mtDNA led to glycerol growth defects similar to those in Δirc3 cells harboring wild-type mtDNA (Fig. S1E). These results are suggestive of a role for IRC3 independent of regulating RNA splicing. The observed reduction in steady-state levels of COX1, ATP8, and ATP6 could be due either to a defect in polycistronic transcript formation or in its subsequent processing. The COX1, ATP8, and ATP6 genes are transcribed on a polycistronic RNA that undergoes processing at two unique maturation sites downstream of COX1 3′ UTR to yield an independent ATP8/ATP6 transcript (53, 54). In order to test whether processing was defective, we analyzed the presence of polycistronic mRNA (COX1-ATP8/ATP6) in IRC3 and Δirc3ρ⁺ cells using primers upstream and downstream of these maturation sites (Fig. S1F). We found a reduction in the polycistronic mRNA levels in Δirc3ρ⁺ cells compared to that in wild-type cells consistent with a defect in transcript formation rather than maturation (Fig. 1D).

To ascertain whether mitochondrial translation is also perturbed in Δirc3ρ⁺ cells, newly synthesized polypeptides were labeled in Δirc3ρ⁺ cells either episomally expressing IRC3 or transformed with vector backbone after growth in glucose and galactose. Growth of cells in glucose leads to glucose repression, whereby expression of genes required for alternate carbon sources as well as mitochondrial oxidative phosphorylation is downregulated. However, growth in galactose does not repress genes required for either fermentation or cellular respiration and thus considered to be neutral for these pathways (55–57). De novo labeling of mitochondrial proteins in Δirc3ρ⁺ cells grown on glucose indicated a severe defect in incorporation of ³⁵S-labeled methionine and cysteine into newly synthesized proteins, in comparison to the wild type (Fig. 1F). In contrast, de novo labeling of mitochondrial proteins in Δirc3ρ⁺ cells grown in galactose led to reduced incorporation of ³⁵S-labeled methionine and cysteine into newly synthesized Cox1 polypeptides in comparison to wild-type cells (Fig. 1F). This indicates that Irc3p is pivotal for mitochondrial translation irrespective of the type of carbon source in which the cells are grown.

Within the mitochondria, transcription and translation are tightly coupled, presumably to make mitochondrial gene expression more efficient and faster (58–61). Thus, diminished accumulation of mRNA levels in Δirc3ρ⁺ cells could be indirect consequences of altered translation rates or vice versa.

IRC3 is essential for mitochondrial translation during basal OXPHOS activity.

To ascertain whether transcription or translation is primarily controlled by IRC3, two conditional alleles of IRC3 designated irc3-1 and irc3-2 were generated as described in Materials and Methods and further characterized. The altered amino acids are described in Table S3. Conditional alleles allowed us to examine the consequences on mitochondrial transcripts and newly synthesized protein products upon rapid depletion of active pools of Irc3 protein (see below).

Cells expressing either temperature-sensitive alleles or the wild-type allele of IRC3 when cultured in glucose at the permissive temperature of 23°C gave rise to progeny that was able to utilize glycerol at the same rate. On the other hand, cells expressing either irc3-1 or irc3-2 when subcultured in glucose at the nonpermissive temperature of 37°C gave rise to fewer progeny that could utilize glycerol (Fig. 2A). We confirmed that cells expressing either irc3-1 or irc3-2 retained mtDNA copy numbers similar to wild-type levels upon subculturing in glucose for 24 h at both at 23°C and 37°C (Fig. 2B; Fig. S2A).

FIG 2.

Rapid depletion of functional IRC3 in cells grown in glucose leads to aberrant mitochondrial translation. (A) Tenfold serial dilutions of glucose-grown Δirc3ρ⁺ cells expressing either an episomal copy of IRC3, vector, irc3-1, or irc3-2, spotted on glucose and glycerol plates, and incubated at 23°C and 37°C are shown. (B) The ratio of nuclear to mitochondrial DNA was estimated in the indicated strains after subculturing in glucose for the indicated time period at 23°C or 37°C. (C) Transcript levels of mitochondrial genes were assayed in the indicated strains grown in glucose. 21S rRNA and 15S rRNA levels were detected by EtBr. Samples were normalized for total RNA. (D) Mitochondrial protein synthesis was measured in the indicated strains grown in glucose at either 23°C or after shift to 37°C for 15 min. Labeling reaction was carried out by incorporation of ³⁵S-labeled methionine and cysteine in the presence of cycloheximide to inhibit cytosolic protein synthesis at 23°C or 37°C for 30 min. Mitochondria from labeled cells were isolated, and equivalent mitochondrial proteins were separated on a 17.5% SDS-PAGE gel. Radiolabeled proteins were visualized by phosphorimaging. At the bottom is a Coomassie blue-stained gel of radiolabeled mitochondrial protein extracts. (E) Mitochondrial ribosomal subunits from the indicated strains cultured in glucose at 37°C for 4 h were separated on a 10-to-30% sucrose gradient. RNA content was monitored by measuring the absorbance at 254 nm. The positions of the 37S and 54S peaks are labeled. The relative ratio of the 54S/37S peaks is listed below each profile.

To determine whether IRC3 primarily regulates transcription, we measured the levels of mature transcripts by using reverse transcriptase PCR in cells harboring either IRC3, irc3-1, or irc3-2 cultured at either 23°C or after a shift to 37°C for 2 h. We found that levels of mitochondrial mRNA in irc3-1 and irc3-2 cells were similar to those in wild-type cells at 23°C or after shift to 37°C for 2 h (Fig. 2C). These results indicate that reduced steady-state levels of mitochondrial mRNA in the Δirc3ρ⁺ strain in glucose-grown cells (Fig. 2C and 1D) could be an indirect consequence of the loss of active pools of Irc3.

To determine whether Irc3p regulates mitochondrial translation, incorporation of ³⁵S-labeled methionine and cysteine into newly synthesized mitochondrial proteins was analyzed in IRC3, irc3-1, and irc3-2 cells grown in glucose, galactose, and glycerol. Cells expressing IRC3, irc3-1, and irc3-2, when grown on glucose at 23°C, showed similar rates of incorporation of radiolabel into newly synthesized mitochondrial proteins (Fig. 2D). However, when cells were shifted to 37°C for 15 min, cells expressing temperature-sensitive alleles irc3-1 and irc3-2 showed reduced incorporation of ³⁵S-labeled methionine and cysteine in comparison to that of wild-type cells, indicating the involvement of Irc3 in optimal regulation of mitochondrial translation (Fig. 2D and Fig. S2B). Interestingly, incorporation of radiolabel in newly synthesized proteins was compromised to a greater extent in irc3-2 cells than in irc3-1 cells, which could be due to mutation at a different site (Table S3). These results indicate that synthesis of all mitochondrially encoded proteins is affected early upon depletion of the active pools of Irc3 in cells under growth conditions containing glucose as the carbon source.

To rule out the possibility that the protein translation defects observed in irc3-1 or irc3-2 mutant cells at 37°C were due to an indirect consequence of aberrant mitochondrial ribosomal assembly/stability, the mitochondrial ribosomes isolated from glucose-grown cultures of IRC3, Δirc3ρ⁺, irc3-1, and irc3-2 cells after shift to 37°C for 4 h were separated on a sucrose density gradient. The relative ratios of large mitochondrial subunits (54S) to small mitochondrial ribosomal subunits (37S) and their levels were quantified and compared. The levels of mitochondrial small ribosomal subunits (37S) in Δirc3ρ⁺ cells were severely reduced, leading to altered 54S/37S ratios, whereas in wild-type, irc3-1, and irc3-2 cells, similar 54S/37S subunit ratios and levels were observed upon a shift to 37°C for 4 h (Fig. 2E). This indicated that altered ratios in Δirc3ρ⁺ cells are an indirect consequence of disrupted mitochondrial function due to the absence of Irc3 rather than from a direct role for Irc3 to promote ribosome assembly. Since 54S/37S ratios are unchanged in irc3-1, irc3-2, and IRC3 cells at 37°C, this indicates that the defective mitochondrial protein synthesis observed in irc3-1 and irc3-2 cells at 37°C is not an indirect consequence of defective mitochondrial ribosomal assembly/stability.

We have further observed that mitochondrial translation is regulated by IRC3 differentially depending upon the available carbon source. Equivalent levels of newly synthesized mitochondrial proteins were observed in IRC3, irc3-1, and irc3-2 cells that had been cultured in glycerol or galactose at 23°C or followed by incubation at 37°C for 2 h (Fig. S2C and S2D). Consistently, cells expressing IRC3, irc3-1, and irc3-2 when cultured in glycerol and galactose at 23°C or 37°C gave rise to progeny that were able to utilize glycerol to the same extent (Fig. S2E and S2F). Reduction in levels of de novo Cox1p in Δirc3ρ⁺ cells (Fig. 1F) is unlikely to be due to defects in ribosome assembly/levels, as the 54S/37S ratios were similar in Δirc3ρ⁺ and IRC3 cells cultured in glycerol for 12 h at 30°C (Fig. S4). Hence, taken together, our data support a role for Irc3 in regulating mitochondrial translation in response to metabolic cues.

Irc3p fractionates with small mitochondrial ribosomal subunits independent of metabolic cues.

Given that IRC3 is a predicted RNA helicase and a reduction in functional IRC3 in cells leads to aberrant mitochondrial protein synthesis suggest a role for Irc3p in mitochondrial translation initiation and elongation, either of which would require Irc3p to be associated with the mitochondrial ribosome. In the absence of antibodies to the endogenous protein, to detect Irc3 within the mitochondria, we tagged IRC3 with 13×myc at the C terminus. We confirmed that IRC3-myc encodes a functional protein localized to the mitochondria (Fig. S3A and S3B). Interestingly, we had observed that mitochondrial translation was perturbed to different levels in cells expressing IRC3 mutants when cultured under conditions that maintain mitochondrial OXPHOS function at basal or elevated levels (Fig. 1F and 2D; Fig. S2B and S2C). Thus, we determined whether Irc3-myc physically associates with mitochondrial ribosomes in exponentially growing cells cultured in glucose and glycerol. In addition, we examined the association of Irc3-myc with mitochondrial ribosomes in cells that had attained stationary phase. Mitochondrial ribosomes were separated on a 10-to-30% sucrose gradient as described in Materials and Methods, and individual fractions were probed for the presence of Mrp7 and Mrp13 to ascertain the fractions containing the 37S small subunit and 54S large subunit. Irc3-myc peaked in fractions containing Mrp13 in mitochondria from exponentially growing cells cultured in glycerol and glucose (Fig. 3A). A small proportion of the total signal from Irc3-myc also cofractionated with both Mrp7 and Mrp13, indicating an association with a 74S subunit in mitochondria from exponentially growing cells cultured in glucose and glycerol (Fig. 3A). In contrast, when mitochondrial ribosomes were separated on a sucrose gradient from stationary-phase cells cultured in glycerol, the majority of Irc3-myc was found at the top of the gradient and only a small fraction of the Irc3-myc pool was present in fractions containing Mrp7, followed by a minor pool in heavier fractions (Fig. 3B).

FIG 3.

Irc3p associates with small subunits of mitochondrial ribosome during exponential growth phase. Mitochondrial ribosomes from cells expressing functional IRC3-Myc, which are cultured either in glucose or glycerol to exponential phase (A) or in glycerol to stationary phase (B) were separated by ultracentrifugation on a 10-to-30% sucrose gradient containing 400 mM NH₄Cl at 135,000 × g for 4 h. Equivalent protein fractions were TCA precipitated, separated on an SDS-PAGE gel, and subjected to immunoblot analysis using antibodies against c-Myc, Mrp7p, and Mrp13p. The positions of 37S, 54S, and 74S are shown at the bottom. The uppermost fraction of sucrose gradient upon centrifugation is labeled as L. An asterisk indicates a cross-reacting signal obtained from anti-Myc antibody likely to be a cleavage product of Irc3 during sample processing.

Irc3p is involved in mitochondrial translation elongation.

The absence of or reduction in active pools of Irc3 led to reduced mitochondrial protein synthesis. This could be due to a defect in either translation initiation or elongation or both. To directly examine the role of IRC3 in translation initiation from mitochondrial mRNA, we used engineered strains YC162, RG140, and RG139, carrying a mitochondrial reporter gene in Δirc3 cells expressing either IRC3, vector, irc3-1, or irc3-2. In these strains, arginine 8 (Arg8), an essential matrix-localized enzyme required for arginine biosynthesis, has been deleted from the nuclear genome and a recoded version termed Arg8^m is expressed from the mitochondrial DNA, specifically replacing the coding sequences of Cox1, Cox2, and Cox3 and keeping the 5′ and 3′ UTRs intact (Fig. S5) (62). Functional mitochondrial translation initiation in these strains can be scored based on the ability of cells to grow on synthetic plates lacking arginine (SD-Arg; where SD is 0.67% yeast nitrogen base without amino acids containing 2% glucose). YC162, RG140, and RG139 deleted for IRC3 showed a growth defect on SD-Arg with reduced Arg8^m protein accumulation in mitochondria (Fig. 4A and B), consistent with multiple defects in mitochondrial gene expression (Fig. 1). However, engineered Δirc3 cells episomally expressing either the irc3-1 and irc3-2 allele showed growth rates similar to that of wild-type cells on SD-Arg at both permissive and nonpermissive temperatures (Fig. 4A). Consistently, steady-state levels of Arg8^m in mitochondria from cells cultured at a nonpermissive temperature were also similar in these cells (Fig. 4B). These results indicate that although IRC3 is a key regulator of mitochondrial translation, defects observed after rapid depletion of active pools of Irc3 do not severely compromise the initiation process.

FIG 4.

Irc3p regulates mitochondrial translation elongation and not initiation. (A) Serial dilutions of Δirc3ρ⁺cells expressing either a wild copy of IRC3, vector, irc3-1, or irc3-2 harboring mtDNA modified to carry Arg8^m reporter cassettes in place of COX1, COX2, and COX3 open reading frames were cultured in glucose prior to spotting on glucose and synthetic medium lacking arginine and incubated at the indicated temperature. (B) Steady-state levels of mitochondrial Arg8^m were measured by immunoblot analysis in the indicated strains at the nonpermissive temperature. Anti-F1β was used as a loading control. (C) Serial dilution of Δirc3ρ⁺ cells expressing either a wild-type copy of IRC3, vector, irc3-1, or irc3-2 harboring mtDNA expresses Arg8^m reporter cassette in place of COX1 ORF, while COX1 and COX2 ORF were placed under COX2 5′ and 3′ UTRs cultured in glucose prior to spotting on glucose, glycerol, and synthetic media lacking arginine and incubated at indicated temperatures. (D) Mitochondrial protein synthesis was measured in the indicated strains upon shift from 23°C to 37°C for 30 min. Labeling reaction was carried out by incorporation of ³⁵S-labeled methionine and cysteine in the presence of cycloheximide to inhibit cytosolic protein synthesis at 37°C for 30 min. Mitochondria from labeled cells were isolated, and equivalent amounts of proteins were separated on a 17.5% SDS-PAGE gel. Radiolabeled proteins were visualized by phosphorimaging. At the bottom is a Coomassie blue-stained gel of radiolabeled mitochondrial protein extracts. (E) Steady-state levels of mitochondrial Arg8^m and F1β were measured in the indicated strains cultured at 37°C.

To differentiate the role of IRC3 more specifically in mitochondrial translation initiation and elongation, we used engineered strains XPM171a and XPM78a with a mitochondrial reporter gene in Δirc3ρ⁺ cells expressing either IRC3, vector, irc3-1, or irc3-2. In XPM171a, the recoded Arg8^m gene was placed under the control of COX1 5′ and 3′ UTRs, while COX1 and COX2 open reading frames (ORFs) were placed under the control of COX2 5′ and 3′ UTRs (Fig. S5) (27). This allows translation initiation to be monitored as a function of growth on SD-Arg and translation elongation to be scored as a function of growth on glycerol (YPG [1% yeast extract, 2% peptone, 3% glycerol]). When cells were cultured in glucose medium, Δirc3 cells were severely defective for growth on SD-Arg and YPG in comparison to wild-type cells (Fig. 4C), consistent with compromised mitochondrial gene expression (Fig. 1). In contrast, Δirc3 cells episomally expressing either the irc3-1 or irc3-2 allele were defective for growth on YPG but not SD-Arg at 37°C in comparison to IRC3 cells, while at 23°C, IRC3, irc3-1, and irc3-2 cells grew at similar rates on both SD-Arg and YPG (Fig. 4C). Consistent with a defect in translation elongation in glucose-cultured cells, we found that incorporation of ³⁵S-labeled methionine and cysteine into newly synthesized Cox1 as well as other mitochondrial encoded proteins were reduced in Δirc3ρ⁺, irc3-1, and irc3-2 cells upon shift to 37°C from 23°C for 30 min (Fig. 4D). Consistent with no significant defects in translation initiation, Δirc3ρ⁺ cells episomally expressing either IRC3, irc3-1, or irc3-2 at 37°C accumulated similar levels of Arg8^m in the mitochondria (Fig. 4E).

XPM78a expresses 512 nucleotides of intronless COX1 fused to Arg8^m under the control of COX1 5′ and 3′ UTRs (Fig. S5A). Within the Arg8^m sequence, the cleavage site for pre-Arg8 is present, such that Cox1 and Arg8^m function independently (27). This allows translation elongation to be scored on SD-Arg and glycerol plates. Cells harboring the Δirc3 allele were compromised for growth on both SD-Arg and glycerol in comparison to the wild type (Fig. S5B). Consistently, incorporation of ³⁵S-labeled methionine and cysteine into newly synthesized Cox1 and Arg8^m and steady-state Arg8 levels in Δirc3 cells were reduced in comparison to that of wild-type cells (Fig. S5C and S5D). Growth defects in Δirc3ρ⁺cells expressing irc3-1 and irc3-2 on SD-Arg and glycerol plates could not be ascertained, as there was a drastic loss of cell viability at the nonpermissive temperature when the cells were incubated for more than 6 h (data not shown). In addition, we were not able to visualize the Cox1-Arg8 fusion product, likely due to its labile nature. Taken together, these results indicate that Irc3p is involved in mitochondrial translation elongation.

Rho suppressors of the Δirc3 strain restores translation differentially based on metabolic cues.

We have shown that Irc3p regulates mitochondrial translation differently in cells based upon the carbon source available for growth. Identification of suppressors of deletion strains often leads to revelation of novel genetic networks in play. Thus, we isolated spontaneous Δirc3ρ⁺ suppressors by using two independent strategies. In the first screen, Δirc3ρ⁺ suppressors were identified by plating the Δirc3ρ⁺ strain for single-cell colonies on solid medium containing glycerol. Cells that were able to utilize glycerol at a higher rate were initially scored as suppressors. Two Δirc3 suppressors (Δirc3ρ^SUP1 and Δirc3ρ^SUP2) that were able to maintain growth in respiratory medium after repeated subculturing in glucose were analyzed further (Fig. 5A). A second independent screen used the Δirc3ρ⁺ strain carrying an unlinked ade2 mutation in the background. Strains carrying an ade2 mutation are red in color when grown in glucose as long as their mitochondrial function is intact, as are strains carrying IRC3 and ade2 (Fig. 5A). Strains with compromised mitochondrial function that are deficient for respiration are white in color (63, 64), as are, e.g., Δirc3,ade2 strains (Fig. 5A). To find second-site Δirc3 suppressors, Δirc3, ade2 cells were plated on solid glucose medium and scored for red colony sectors. Four suppressors (Δirc3,SUP3ρ⁺, Δirc3ρ^SUP4, Δirc3ρ^SUP5, and Δirc3ρ^SUP6) were identified and further confirmed for growth on glycerol medium (Fig. 5A). Interestingly, the suppressors identified by the first screen formed white colonies on glucose plates in spite of the presence of an ade2 mutation, and on glycerol plates, the colony size was larger than that of the wild type (Fig. 5A).

FIG 5.

Restoration of mitochondrial protein synthesis in Δirc3ρ⁺ suppressors is dependent on the carbon source used for cell growth. (A) Tenfold serial dilutions of the wild type and Δirc3ρ⁺and Δirc3ρ⁺ suppressors spotted on glucose and glycerol plates are shown. (B) Mitochondrial protein synthesis was measured in wild-type, Δirc3ρ⁺, and the indicated Δirc3ρ⁺ suppressor strains cultured in glucose and glycerol at 30°C. Labeling reaction was carried out by incorporation of ³⁵S-labeled methionine and cysteine in the presence of cycloheximide to inhibit cytosolic protein synthesis at 30°C for 30 min. Mitochondria from labeled cells were isolated, and proteins were separated on a 17.5% SDS-PAGE gel. Radiolabeled proteins were visualized by autoradiography. At the bottom is a Coomassie blue-stained gel of radiolabeled mitochondrial extracts. (C) Transcript levels of the indicated mitochondrial genes were assayed in wild-type, Δirc3ρ⁺, Δirc3ρ^SUP1, and Δirc3ρ^SUP2 cells under glucose culture conditions. 21S rRNA and 15S rRNA levels were detected by EtBr staining.

Diploids generated by crossing the suppressor strains with the Δirc3ρ⁰ strain retained the ability to grow on glycerol, thereby indicating that the suppressor mutation was dominant in character. Among the six suppressors identified, five suppressors contained a mutation in the mitochondrial genome and one suppressor contained a mutation in the nuclear genome, as determined based on the crosses described (Fig. S6). As we have shown before that Irc3 regulates mitochondrial translation differentially in glucose-grown cultures and galactose/glycerol-grown cultures, we determined whether mitochondrial translation was restored to the same levels in all suppressors cultured in glucose or glycerol. We found that incorporation of ³⁵S-labeled methionine and cysteine into newly synthesized mitochondrial proteins were restored to wild-type levels in Δirc3,SUP3ρ⁺, Δirc3ρ^SUP4, Δirc3ρ^SUP5, and Δirc3ρ^SUP6 strains independent of carbon source in the growth medium (Fig. 5B). Interestingly, in Δirc3ρ^SUP1and Δirc3ρ^SUP2 strains, mitochondrial translation was restored to nearly wild-type levels when glycerol was used as a carbon source in growth medium. However, mitochondrial translation was not restored to wild-type levels when glucose was used as a carbon source in growth medium (Fig. 5B). Levels of mitochondrial mRNA were similar in Δirc3ρ^SUP1 and Δirc3ρ^SUP2 suppressors and in wild-type cells grown in glucose, indicating that reduced translation levels in glucose-grown cells were not due to reduced transcript levels of mitochondrial mRNA (Fig. 5C). These results further indicate that Irc3 regulates mitochondrial translation and that the mode of regulation is dependent on the carbon sources available for cell growth.

DISCUSSION

Translation is a multistep process involving initiation and peptide elongation followed by termination and ribosome recycling. Translation rates are regulated by numerous protein factors and mRNA secondary structures that vary with environmental conditions (65). In mitochondria, the translation system has an added layer of complexity, where the machinery is tightly associated by the mitochondrial inner membrane and requires membrane-embedded mRNA-specific translation activator for protein synthesis (4). These presumably are the reasons why a membrane-free in vitro translation system has not yet been developed, and the mechanism underlying mitochondrial protein synthesis remains poorly understood.

In this study, we used genetic and biochemical methods to investigate the role of the putative DExH/D helicase gene IRC3, which could be speculated to modify structures involving either DNA-DNA or RNA-RNA or DNA-RNA hybrids during mitochondrial DNA synthesis/repair and gene expression, respectively. We observed that cells deleted for IRC3 rapidly lose the ability to utilize glycerol as the sole carbon source, followed by a relatively low rate of mitochondrial DNA loss (Fig. 1), consistent with previously reported examples of proteins required for mitochondrial gene expression (43, 49, 50). Furthermore, we demonstrate as expected for a protein to regulate mitochondrial translation, Irc3-myc cofractionates with mitochondrial small ribosomal subunits (Fig. 3A), deletion of IRC3 led to decreased mitochondrial translation (Fig. 1), and Δirc3ρ⁺ suppressors restored mitochondrial translation to distinct levels in response to metabolic cues (Fig. 5). Although the precise molecular details by which Irc3 regulates mitochondrial translation is unknown, our results support a model where Irc3 regulates translation elongation. First, upon direct examination of mitochondrial translation in irc3-1 and irc3-2 cells, we observed a decrease in de novo polypeptide synthesis upon shift to the nonpermissive conditions within the first 15 min (Fig. 2). Second, using the Arg8^m reporter gene, we observed that Irc3 is required specifically for translation elongation (Fig. 4; see Fig. S5 in the supplemental material).

It has been well established that in response to metabolic cues (glucose versus galactose/glycerol), yeast cells differentially regulate gene expression (66–70). Culturing cells in glucose, galactose, or glycerol leads to differential utilization of mitochondrial OXPHOS for ATP production. In glucose, mitochondrial OXPHOS activity is maintained at basal levels, while in galactose or glycerol, this activity is elevated (71, 72). A shift of yeast cells from glucose to galactose/glycerol leads to increased translation of nuclear and mitochondrially encoded OXPHOS subunits. Increased mitochondrial translation is brought about as a consequence of upregulated expression of nucleus-encoded mRNA-specific translation activators (22, 23, 73). Could additional molecular factors with novel mechanisms that modulate mitochondrial translation in response to carbon source exist? We found that Irc3 differentially regulates mitochondrial translation when cells are cultured under conditions which require OXPHOS at basal versus elevated levels (Fig. 1 and 2; Fig. S2). When glucose is used as the carbon source, protein translation in Δirc3ρ⁺ mitochondria is defective for all transcripts, whereas when galactose is used as the carbon source, protein translation from Cox1 mRNA is predominantly reduced. This contrasts with the observation that accumulation of mitochondrial transcripts was disrupted to the same extent irrespective of carbon source used for growth (Fig. 1). When irc3-1 and irc3-2 cells were cultured in glucose at restrictive temperatures, the rates of protein translation mirrored protein translation rates from Δirc3ρ⁺ cells albeit to a less severe level. This could be due to expression of partially functional proteins from irc3-1 and irc3-2 alleles that are distinctly labile at elevated temperatures. However, when irc3-1 and irc3-2 cells were cultured in galactose or glycerol at restrictive temperatures, mitochondrial translation was not altered, and their progeny cells were not defective for growth on glycerol. Coincidentally, our screening for temperature-sensitive mutants was biased toward disrupting IRC3 function specifically in glucose-cultured cells. Thus, taken together, these findings lead one to speculate that Irc3p differentially regulates mitochondrial translation in response to metabolic cues.

Genetic suppressors have been routinely used to understand cellular networks that play intertwined roles. We screened for Δirc3ρ⁺suppressors, based on the ability to restore normal mitochondrial function in either glucose or glycerol. Intriguingly, Δirc3ρ⁺ suppressors that were isolated restored mitochondrial translation to different extents based on the carbon source added to the growth medium. Δirc3ρ⁺ suppressors that were isolated for restoration of mitochondrial function on glycerol restored mitochondrial translation to wild-type levels only when glycerol and not glucose was the carbon source added to the growth medium. In contrast, Δirc3ρ⁺suppressors that were isolated for restoration of mitochondrial function on glucose restored mitochondrial translation to the same extent irrespective of whether glucose or glycerol was provided for cell growth (Fig. 5). Although it is not possible to comment on the exact nature of the suppressor mutation(s), it leads to increased processivity of the translation machinery. A mutation(s) on mtDNA in glucose-derived suppressors likely increases the translation processivity to a greater extent than that in glycerol-derived suppressors. Could this difference in translation processivity in Δirc3ρ⁺suppressors hypothetically indicate the existence of altered mRNA structures of various degrees of compactness in mitochondrial transcripts between cells grown in glucose versus glycerol? It is not completely understood how mRNAs fold into compact structures in the cell. However, secondary and tertiary structures in the 5′ UTR, coding sequence (CDS), and 3′ UTR are well described regulators of essential processes such as translation initiation, translation elongation, ribosomal traffic jams, and mRNA stability (reviewed extensively in references 74 to 77). Ribosome processivity during the process of translation elongation is believed to be the principal mechanism that leads to its secondary-structure resolution within the CDS (78–80). Interestingly, in glucose-grown yeast cells, mitochondrial transcripts are abundant, although their translation is at the basal level. Upon a shift to glycerol, increases in individual mRNA translation rates are greater than increases in their transcription rates (22, 81). Thus, one could imagine that compact mRNA molecules contribute to the previously described basal translation rates in glucose. Conversely, when glycerol is the carbon source for cell growth, less-compact mitochondrial mRNA structures exist, possibly because of the enhanced level of association with the ribosomes, requiring Irc3 for translation of a subset of mRNA such as for Cox1.

How could Irc3p differentially regulate mitochondrial translation when associated with mitochondrial small ribosomal subunits in both glucose and glycerol? Two nonexclusive hypothetical models could be invoked to address this question, both of which require Irc3‘s association with different interacting partners. These interacting partners could be proteins whose expression is controlled by metabolic cues, such as those that have been described as a part of large-scale proteome remodeling that occurs when cells are shifted from glucose to galactose/glycerol (22). In the first model, these protein partners in association with Irc3 target the regions of mRNA with compact local secondary RNA structures that are maintained in the mitochondria as a function of their growth condition and which require unwinding during translation elongation. An alternate model could be that these distinct interacting partners for Irc3 are associated with the compact mRNA structures. During the process of translation elongation, Irc3 bound to the small subunit is responsible for displacing these proteins, thereby melting or unwinding the structures required for translation elongation. Future experiments targeted to parsing the mechanism by which Irc3 regulates mitochondrial translation will require these hypothetical models to be accounted for.

Deletion of IRC3 has previously been reported to accumulate double-stranded mitochondrial DNA breaks (47). ATPase activity of the Irc3 protein is essential in vivo and is stimulated in vitro by DNA molecules, especially those mimicking recombination intermediates (47, 48, 82). Here, we have clearly shown that IRC3 is required for mitochondrial translation in vivo. It is not well understood why disruption of mitochondrial translation often leads to loss of mitochondrial DNA. Is it possible that SFII DExH helicases can target both mtDNA and mitochondrial mRNA, thereby forming a link between optimal translation and mtDNA stability? Among the four annotated SFII DExH helicases, SUV3, which is the RNA helicase component of the mitochondrial exoribonuclease complex in addition to having a role in RNA turnover (44, 46), has also been implicated in mitochondrial genome stability (83–85). A similar link between ribosome biogenesis with DNA replication has been found in both bacteria and yeast. In bacteria, Obg GTPases, which are known to control large ribosomal subunit biogenesis (86, 87), link essential processes required for cell cycle progression, such as DNA replication, by a yet unknown mechanism (88–91). Nog1, the nucleolar member of the Obg family of GTPases in yeast which is required for 60S ribosomal subunit biogenesis (92, 93), is also associated with the origin recognition complex in yeast (94, 95). Interestingly, we found Irc3 to predominantly cofractionate with mitochondrial ribosomes in cells that are in logarithmic growth phase. However, in cells that have entered stationary phase, only a small fraction of Irc3 is cofractionated with mitochondrial ribosomes (Fig. 3). It is interesting to speculate that Irc3 during rapid growth is involved in the regulation of mitochondrial translation and switches to mtDNA maintenance at other times, thus establishing a link between these two processes.

Mitochondrial functioning vis-à-vis ATP generation is altered depending on whether the cell is poised for rapid proliferation or in a quiescent/differentiated state. Proliferating cells, including pluripotent stem cells, use glycolysis even in the presence of oxygen, whereas differentiated cells use mitochondrial oxidative phosphorylation (96–98). Saccharomyces cerevisiae shows a similar plasticity in response to metabolic cues; i.e., when grown aerobically in glucose, yeast cells derive ATP from fermentation while producing ethanol as a by-product. Under these conditions, the metabolism in yeast cells is akin to proliferating mammalian cells. Upon exhaustion of glucose, yeast cells switch to oxidative phosphorylation for ATP generation from the ethanol produced during fermentation to sustain itself, akin to differentiated cells (99, 100). We have shown that IRC3 regulates mitochondrial translation distinctively in yeast mitochondria in response to metabolic cues. The closest orthologue for IRC3 in humans based on BLAST is an uncharacterized putative RNA helicase (GenBank accession no. BAG50877) that shares 24% identity and 39% similarity within the helicase core only and is predicted to localize to mitochondria with 84% probability (101). Thus, although a clear orthologue for IRC3 is not present in mammalian cells, one could speculate that a similar mechanism exists in mammalian mitochondria to control synthesis of the OXPHOS protein complex whereby energy demands and signals governing the cellular state are integrated.

MATERIALS AND METHODS

Yeast strain, primers, and media.

The yeast strains used are listed in Table S1 in the supplemental material. The sequences of primers used are listed in Table S2. The complete media used were YEP (1% yeast extract and 2% peptone) containing 2% glucose (YPD), 2% galactose (YPGal), or 3% glycerol (YPG) as a carbon source. Synthetic minimal media (0.67% yeast nitrogen base without amino acids containing 2% glucose [SD], 2% galactose [SGal], 3% glycerol [SG], 0.1% 5-fluoro-orotic acid and 2% glucose (5FOAD), or 0.006% canavanine sulfate and 2% glucose [SCAN]) were supplemented with the appropriate amino acids when required, as described previously (102).

Generation of plasmids and strains.

To generate an episomally expressed IRC3, IRC3 with 634 bp upstream of the start codon and 1 kb downstream of the stop codon in addition to the ORF was PCR amplified using the YDR332wup and YDR332wdn primer pair (Table S2) and cloned into pRS426 and pRS315 at the NotI site to generate pRS426:IRC3 (pJk20) and pRS315:IRC3 (pJk22). In order to generate a construct expressing IRC3 with C-terminally tagged Myc under its endogenous promoter pRS315::IRC3-Myc, Myc-KanMX6 was amplified from the pFA6a-13Myc-kanMX6 plasmid (103) by using the P3 and P4 primer pair (Table S2) and cloned into a pGEM-T easy vector at the TA cloning site to generate pJK55 (KDB784). Myc-KanMX6 from pJK55 was subcloned into pJK34 (IRC3::3HA) as a BamHI and SalI fragment to generate pJK52 (KDB698).

Cells deleted for IRC3 were generated in a diploid strain (KDY335) created by crossing CRY1 (MATα ura3-52 trp1Δ2 leu2-3_112 his3-11 ade2-1 can1-100) and BY4742 (Matα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0). Disruption of IRC3 was carried out by using a PCR-based gene replacement approach, in which the Schizosaccharomyces pombe HIS5 gene was amplified from pFAa-His3MX6 (103) with upstream and downstream regions of IRC3 by using PCR using oligonucleotides F1 and R1 (Table S2). The linear PCR product was transformed into a diploid strain (KDY335) to create a heterozygous diploid strain (KDY394). A plasmid-borne IRC3 (pJK20) was transformed into a heterozygous diploid strain (KDY394) and sporulated, and haploid spores were selected on SCAN to generate a strain with a chromosomal copy of IRC3 deleted and Irc3 expressed episomally with intact mitochondrial DNA (KDY495). In order to eliminate the chance of creating a mutation at a second site in the genome during generation of Δirc3::HIS5 and to possess mtDNA from an isogenic strain, KDY495 was backcrossed six times with BY4741. The haploid strain generated upon backcrossing was named KDY1146. Replacement of the wild-type copy of IRC3 with HIS5 was similarly carried out in YC162, RG139, RG140, and W303ρI⁰ to generate KDY1575, KDY1529, KDY1531, and KDY1589, respectively.

In order to study translation initiation and elongation, the Δirc3 strain carrying the ARG8^m reporter cassette was created by crossing KDY1146 ρ⁰ with XPM171a and XPM78a, respectively. Diploid cells were sporulated, and haploid spores were selected on SCAN, lacking arginine and histidine. To ensure that the arginine gene from the nuclear genome in the haploids is disrupted, mitochondrial DNA was removed by growing cells in YPD supplemented with 25 μg/ml ethidium bromide (EtBr) and tested for growth on SD plates lacking arginine. The corresponding cells unable to grow on SD plates lacking arginine were selected from YPD plates, and the absence of the IRC3 gene from the nuclear background was confirmed by PCR using YDR332wup and YDR332wdn/R1 primers (Table S2) to generate KDY1447 and KDY1540.

Presence of mitochondrial DNA.

The presence of mitochondrial DNA in Δirc3 cells expressing either IRC3 (KDY1355) or empty vector (KDY1366) was measured by crossing cells with IRC3 ρ⁰ testers, followed by selection of diploid cells. The percentage of diploid cells that were able to utilize both glycerol and glucose was indicative of the presence of mtDNA in the haploid strain.

To measure the mitochondrial DNA copy number with respect to nuclear DNA in Δirc3 cells expressing either IRC3 (KDY1355) or empty vector (KDY1366), total yeast DNA was isolated and quantitative PCR (qPCR) was performed using primers COX2L and COX2R for Cox2 amplification and ActinF.P and ActinR.P for actin amplification (Table S2). The ratio of signal for Cox2 to actin was used as a measure of the ratio of mitochondrial DNA to nuclear DNA, respectively.

Generation of temperature-sensitive (ts) allele of IRC3.

Temperature-sensitive irc3 alleles were created using error-prone PCR as described previously (104). The coding region was amplified by using the tsmutantF.P and YDR332wdn primer pair (Table S2). PCRs were carried out in the presence of a 5:1 ratio of (dTTP + dCTP):(dATP + dGTP) or (dATP + dGTP):(dTTP + dCTP). The concentration of MnCl₂ was varied from 0.3 to 0.5 mM. The resulting mutated PCR fragment was cloned by gap repair into pJK22. Linearized pJK22 (digested with NsiI) and mutagenized PCR product were cotransformed into KDY494. Transformed cells were selected on SD plates lacking leucine and uracil and patched on 5FOAD to negatively select for URA3 helper plasmid expressing the wild-type allele of IRC3. Cells were then patched on glucose plates and incubated at 23°C and 37°C for 2 days. Cells were patched onto glycerol plates and scored for colony formation at the permissive temperature (23°C) but not at the nonpermissive temperature (37°C). Plasmid DNA was isolated from Δirc3 cells harboring putative temperature-sensitive alleles and amplified in Escherichia coli and sequenced. Two confirmed temperature-sensitive alleles of IRC3 (pJK90 [KDB1019] and pJK93 [KDB1061]) were shuttled into KDY1146 and confirmed for growth at 23°C and not 37°C on YPG.

Isolation of mitochondria and separation of mitochondrial ribosomes.

Mitochondria were isolated from cells grown in YPGal, YPD, or YPG at the indicated temperature at an optical density at 600 nm (OD₆₀₀) of 1 or 3. Mitochondria were isolated by a previously described method (105) with slight modifications as described previously (49). Mitochondrial ribosomes were separated by a sucrose density gradient as described previously (106). Mitochondria were resuspended in buffer D (10 mM Tris-Cl, pH 7.4, 10 mM magnesium acetate, 100 mM NH₄Cl, 7 mM β-mercaptoethanol, and 1 mM phenylmethylsulfonyl fluoride [PMSF]) and incubated on ice for 2 h to generate mitoplasts. Mitoplasts were harvested by centrifugation at 12,000 × g for 10 min, resuspended in buffer D containing 2% NP-40, and further incubated on ice for 30 min to allow swelling. Mitochondrial lysis was achieved by 20 strokes with a tight-fitting pestle Dounce homogenizer. The mitochondrial lysate was clarified by centrifugation at 40,000 × g for 25 min. An equivalent number of clarified lysates was loaded onto a 10-to-30% sucrose gradient containing 10 mM Tris-Cl, pH 7.4, 10 mM magnesium acetate, 7 mM β-mercaptoethanol, and 400 mM NH₄Cl. Ribosomal particles were centrifuged at 135,000 × g for 4 h in a Beckman SW41Ti rotor. Equivalent fractions were collected, and RNA contents were measured by UV absorbance at 254 nm using an ISCO continuous-flow cuvette. Protein samples in each fraction were precipitated using 15% trichloroacetic acid (TCA), separated by SDS-PAGE, and subjected to immunoblot analysis.

Analysis of mitochondrial transcripts.

Cells were grown in either YPD or YPGal medium or shifted to YPG medium. Cells were collected by centrifugation at 2,660 × g for 5 min, washed with water, and resuspended in SB3 buffer (50 mM Tris-Cl, pH 8, 10 mM MgCl₂, 3 mM dithiothreitol [DTT], 1 M sorbitol). Spheroplasts were generated by treatment with Zymolyase (2.5 mg/g cell pellet) for 2 h at 37°C. Spheroplasts were collected by centrifugation at 4,000 × g for 5 min and further lysed by 20 strokes of a Dounce homogenizer. Cell lysate was centrifuged at 1,500 × g for 5 min at 4°C to remove cell debris. Supernatant was collected and centrifuged at 17,550 × g for 15 min at 4°C to pellet the mitochondria. Mitochondria were lysed by the addition of TRIzol, followed by chloroform. The aqueous layer containing mitochondrial nucleic acid was separated and precipitated by the addition of 3 M sodium acetate and isopropanol. The pellet was washed twice with 70% ethanol to remove excess salts and resuspended in water. Isolated RNA was further purified by adding 1 volume of 2 M potassium acetate, pH 4.8, incubated on ice for 30 min, and centrifuged at 13,000 rpm and 4°C for 20 min. Precipitated RNA was treated with DNase I (2,000 U/ml) for 30 min at 37°C to remove contamination of nuclear or mtDNA, followed by precipitation with 3 M sodium acetate and isopropanol. Precipitated RNA was washed twice with 70% ethanol and resuspended in water. cDNA using random primers was synthesized from 25 ng/μl of RNA template. Synthesized cDNA was diluted 5 times prior to utilization in semiquantitative PCRs with a 23-cycle extension for mitochondrial genes COX1, COX2, COB, ATP8, ATP9, and ATP6. Semiquantitative PCR with a 40-cycle extension was used to detect COX1-ATP8/ATP6. Primers used are listed in Table S2. For clarity, agarose gel images shown in the figures were color inverted, resulting in dark bands on a white background.

Analysis of mitochondrial protein synthesis.

Cells were grown in SD, SGal, or SG lacking methionine and cysteine to an OD₆₀₀ of 1 at the indicated temperature as described in figures and figure legends of 23°C or 30°C or upon shift to 37°C from 23°C for the indicated time as described in figures and figure legends. Newly synthesized mitochondrial proteins were labeled with 0.1 mCi of ³⁵S-labeled methionine and cysteine (EasyTag Express³⁵S protein labeling mix; Perkin Elmer NEG77200 [specific activity, 1,175 Ci/mmol]) in the presence of cycloheximide for 30 min as described previously (107) at the indicated temperature of 23°C, 30°C, or 37°C. Mitochondria were isolated from labeled cells as described previously (107). Mitochondrial proteins were quantified by the Bradford assay, and equivalent amounts of protein were separated on a 17.5% SDS-PAGE gel. Labeled proteins were transferred to a nitrocellulose membrane and analyzed by a Fla9000 phosphorimager, or the gels were dried and exposed to X-ray film to visualize labeled proteins. Coomassie blue-stained gels were used as a part of normalization of radiolabeled mitochondrial proteins.

Immunoblot analysis.

Proteins were separated on 7.5% or 10% SDS-PAGE gels and subjected to immunoblot analysis. The following antibodies were used: anti-Myc (1:5,000) (MYC.A7; Epitope Biotech, Inc.), Mrp7 (1:2,000) (108), Mrp13 (1:2,000) (109), Arg8p (21), and Mtg2 (49).